How does government get away from 'black swan' IT projects?

You may have heard this story, or one very much like it, before.

Global organisation. Ageing, incompatible and country-specific systems, antiquated network. Decision made to invest in a single - in this case SAP - system in partnership with a provider (Deloitte).

Starting budget: US$5 million. Changes creep in. New interfaces need to be added, customisations made.

Problems quickly scale. Key deadlines and milestones are missed, and budgets fast blow out.

Total cost at end of project: US$192.5 million and one CIO.

This is the early 2000s experience of retail giant Levi Strauss when it tried to upgrade its old and complex global IT network, as detailed in the Harvard Business Review.

It’s a classic example of a “black swan” project, a term coined by Nassim Nicholas Taleb to describe high-impact events that are rare and unpredictable, but in hindsight not all that unexpected.

In Australia, you might consider Queensland Health’s IBM payroll failure, last year’s online Census disaster (also involving IBM), or NSW’s LMBR school systems mess to be local examples of black swan projects.

Project delays and budget blowouts are by no means limited to the public sector, but we hear more about them simply given the nature of the beast.

Two of Australia’s highest-profile CIOs, Gary Sterrenberg from the Department of Human Services and Peter Lawrence from the Department of Defence, are currently staring down the looming outlines of black swans hovering menacingly in the distance.

Making waves

Sterrenberg is helming what is Canberra’s highest-profile IT project at the moment: the billion-dollar “once-in-a-generation” welfare payments infrastructure transformation (WPIT).

Given the green light in April 2015, the project will see the department migrate off its Model 204 mainframe and 30 year-old income support integrated system (ISIS) to an SAP solution, delivered by Accenture.

To reduce some of the risk involved in the highly complex project - the platform processes 50 million payment transactions each day and is based on more than 30 million lines of code - DHS has opted to split the effort up into five tranches, with suppliers bidding for each new bundle of work.

It is a massive undertaking.

Speaking at a Trans Tasman Business Circle event yesterday, Sterrenberg described the steps he is taking with WPIT to try to avoid it becoming a dreaded black swan project.

The speed at which new technologies are being developed and put on the market is one challenge that will need to be overcome to prevent WPIT being slapped with this label, he said.

“New technology is evolving at such a rapid rate that you can’t afford to have a static architecture. You have to have something that’s a lot more dynamic, because by the time you’ve finished developing it, the technology is likely to be out of date,” Sterrenberg said.

“So we have to refactor projects to be a lot more agile than they used to be. You can’t do WPIT in a waterfall way. It’s more likely to be thousands of small projects that deliver incremental change. Technology is driving us to think about project delivery in a very different way.

“This is less of a challenge for the technology side, it’s going to be more of a challenge for the business to absorb those thousand changes into the networks, into the way we have to deal and communicate with customers.”

One way DHS has implemented this approach is by tweaking parts of the system while it goes through initial procurement and design processes.

It’s partly a response to demands by the Australian public on the type of technology it expects to be delivered by the federal government.

“We’re driven largely by the Australian public’s expectation levels [in terms of the technology implemented] at the moment. They’re no longer accepting of anything less than what they get at the banks or at Coles, for example,” Sterrenberg said.

The hope is that delivering smaller, incremental change to users of the system faster while tackling the foundation components in the background will help sell the wider transformation efforts to the public, and generate some goodwill for the effort.

Turning a black swan white

Defence CIO Peter Lawrence is facing a very similar challenge, and he’s got more than one core overhaul to contend with.

Whether it’s the massive next-generation desktop overhaul, mammoth data centre deal, $1 billion ERP transformation, HR struggles, or tumultuous terrestrial communications replacement, Lawrence is no stranger to complex, costly, and contentious IT projects.

The federal government’s first principles review of the Defence department was not kind, and led to a $17 billion cash injection last year to help fix some of the technology problems that had stemmed from the siloed and sprawling Defence operations.

Defence on the whole was given just 24 months to deliver all 75 recommendations of the first principles review.

This “not great” current state of affairs has prompted Lawrence and CTO Mohan Aiyaswami to consider making some “radical shifts” in how the agency does things.

“One of our challenges is ‘do we go through the traditional steps everybody else has gone through with technology evolution or do we take a look at where we are - which is not great in some areas - and make a radical jump, and how do we do that without getting ourselves into challenges [leading to a black swan project],” Lawrence told the Trans Tasman summit.

“We’re at a point where we’ve got some big choices about how we do things over the next few years. And it’s really about how we enable Defence outcomes more efficiently compared to how we’ve done it in the past - by [breaking things down like DHS] and looking at it in a different way.”

One example of how this shift in mindset has materialised is through a similar approach to DHS: the introduction of a scrum agile inspired implementation plan for tackling the recommendations of the first principles review.

Specific action items are broken down into two-week sprints, not just in IT but across the wider organisation. This agile approach forms one of five key planks of a new technology strategy delivered late last year.

Lawrence is also trying to shift the agency away from a technical architecture mindframe to one of an information architecture, centred around business outcomes and how they are delivered.

“We’re trying to balance .. having an architecture around the information we need to drive the business, and how we provide that in a technology sense,” he said.

“That’s the way forward for us.”

Read on to learn the biggest danger of becoming a black swan